HDL Efflux Capacity and Incident Cardiovascular Events

in the Dallas Heart Study

A Rohatgi,* A Khera, JD Berry, EG Givens, CR Ayers, KE Wedin, IJ Neeland, IS Yuhanna, DJ

Rader, JA de Lemos, PW Shaul*corresponding author

Donald W. Reynolds Cardiovascular Clinical Research Center

0%

5%

10%

15%

20%

25%

0 1 2 3 4Years

Niaspan

Placebo

CETPI

A H T D

AIM-HIGH

HPS2

DalcetrapibTorcetrapib

Drug Class

TrialHDL-C change

NIACIN

AIM-HIGH 25%

HPS2-THRIVE

15%

CETPI

Torcetrapib 72%

Dalcetrapib 35%

Boden WE, et al. NEJM 2011;365:2255-67Barter PJ, et al. NEJM 2007;357:2109-22

Schwartz GG, et al. NEJM 2012;367:2089-99HPS2-THRIVE. NEJM 2014;371:203-12

Rader DJ, Hovingh GK. Lancet 2014;384:618-625

HDL Function: Reverse Cholesterol Transport

apoA-I

apoA-I

apoA-I

ABCA1

ABCA1

ABCG1

SRB1

VLDL

Cholesterol efflux

Adorni MP, et al. J Lipid Res 2007;48:2453-62

Macrophage

Liver

Bile/Feces

Khera AV, et al. NEJM 2011;364:127-35

Study Objective

Objective: Perform a systematic evaluation of the epidemiology of cholesterol efflux capacity in a large unselected population free of cardiovascular disease

.

Hypothesis: Cholesterol efflux capacity varies by sex, race and adiposity and is associated with incident cardiovascular events

Baseline

Follow up

9.4 years

Metabolic phenotype

Cholesterol Efflux Capacity

ASCVD (n=132): MI/Stroke

PCI/CABG CV Death

N = 2924

Age = 30-65 (median 42)

57% Women

49% Black

Dallas Heart Study

Lipoprotein composition

Methods

ACAT inhibitor

cAMP

% effluxApoB-depleted plasma

Total

Normalized efflux

% effluxsample

% effluxreference

Cholesterol Efflux

Capacity

Sankaranarayanan S, et al. J Lipid Res 2011;52:2332-40

  NHDL Cholesterol

(mg/dL)Cholesterol Efflux Capacity

(Normalized)

Women 1657 51 [43-61] 0.99 [0.83-1.18]

Men 1267 44 [37-52] 1.01 [0.84-1.20]

p-value   <0.0001 0.1

       

Black 1443 49 [42-60] 0.98 [0.81-1.18]

Non-Black 1481 45 [38-55] 1.01 [0.85-1.19]

p-value   <0.0001 0.002

Cholesterol efflux capacity is expressed as a ratio of percent cholesterol efflux of the sample normalized to the percent cholesterol efflux of a reference sample.

HDL-C and Efflux by Sex and Race

Determinants of HDL-C and Efflux

HDL Cholesterol   Cholesterol Efflux Capacity

Model R2 = 0.35   Model R2 = 0.03

VariableStd beta estimate

  VariableStd beta estimate

Total cholesterol 0.22   Total cholesterol 0.12

Alcohol intake 0.18   HDL-C 0.09

Age 0.13   Male sex 0.05

Exercise dose 0.08   Hypertension 0.04

Black race 0.07   Alcohol intake 0.03

Smoking -0.06   Black race -0.07

Waist to hip ratio -0.08      

Body mass index -0.17      

Male sex -0.23      

Log Triglyceride -0.39      Model R2 reflects the total contribution of the model to the variance in the dependent variable (range 0.0-1.0).

  HDL CholesterolCholesterol

Efflux Capacity

HDL Cholesterol 0.07*

HDL Particle concentration 0.52* 0.15*

HDL Particle size 0.72* 0.02

Total Cholesterol 0.06* 0.15*

LDL Cholesterol -0.15* 0.10*

Triglyceride -0.45* 0.05*

     

Body Mass Index -0.23* -0.02

Waist to Hip Ratio -0.38* 0.02

Truncal Fat (Dexa) -0.20* 0.004

Visceral Fat (MRI) -0.41* 0.03

HOMA-IR -0.29* -0.05*Values are Spearman correlation coefficients. * p<0.05, otherwise p=NSCEC=cholesterol efflux capacity; HOMA-IR: Homeostatic model assessment

Correlations with Lipoprotein/Metabolic Variables

HDL variables

Lipids

Adiposity

Insulin Res

Cholesterol Efflux Capacity and ASCVD Events

0 1 2 3 4 5 6 7 8 90%

2%

4%

6%

8%

10%

Years

Q1

Q2

Q3Q4

Low efflux

Hi efflux

N=132

ASCVD Events:MIStrokePCI/CABGCV death

Log rank p=0.002

ASCVD: atherosclerotic cardiovascular disease

Unadjusted

TRF-adjusted

Unadjusted

TRF-adjusted

TRF + HDL-C

TRF + HDL-P

TRF + HDL-C + HDL-P

HDL cholesterol

Cholesterol Efflux Capacity

0.64 (0.40-1.03)

0.80 (0.47-1.37)

0.44 (0.27-0.73)

0.30 (0.18-0.50)

0.31 (0.18-0.52)

0.34 (0.20-0.56)

0.33 (0.19-0.55)

1.0Hazard Ratio

0.1 10

TRF + HDL-P 1.08 (0.59-1.99)

Hazard Ratio Q4 vs Q1 (95% CI)

HDL-C, Efflux Capacity and ASCVD Events

N=132/2416 for atherosclerotic cardiovascular disease events. Traditional risk factors (TRF) = age, sex, ethnicity, diabetes, hypertension, current smoking, body-

mass index, total cholesterol, log triglycerides, statin use. HDL-C = HDL cholesterol. HDL-P = HDL particle concentration

Q2

Q3

Q4

Q2

Q3

Q4

Primary ASCVD

Hard ASCVD

0.71 [0.46-1.10]

0.42 [0.26-0.68]

0.33 [0.19-0.55]

0.69 [0.39-1.21]

0.46 [0.25-0.84]

0.40 [0.21-0.74]

Hazard Ratio1.00.1 10

84/2402

132/2416 Hazard Ratio (95% CI)

Q1 Referent

Q1 Referent

Primary ASCVD: MI, stroke, coronary revascularization, and CV death Hard ASCVD: MI, stroke, death from MI or stroke. Models adjusted for age, sex, ethnicity, diabetes, hypertension, current smoking, BMI, total

cholesterol, log triglycerides, statin use, HDL-C, and HDL-P.

Efflux Capacity and Hard ASCVD Events

ModelPrimary ASCVD

(N=132/2416)Total CVD

(N=172/2416)

HR (95% CI) HR (95% CI)

Unadjusted 0.75 (0.61-0.91) 0.84 (0.71-0.98)

TRF-adjusted 0.65 (0.53-0.80) 0.75 (0.63-0.89)

TRF + HDL-C 0.66 (0.54-0.81) 0.77 (0.64-0.91)

TRF + HDL-P 0.69 (0.56-0.84) 0.79 (0.67-0.94)

TRF + HDL-C + HDL-P

0.68 (0.55-0.84) 0.79 (0.67-0.94)

Hazard ratios and 95% CIs for 1 SD increase in efflux capacity. Primary ASCVD: MI, stroke, coronary revascularization, and CV deathTotal CVD: ASCVD + peripheral revascularization, CHF and afib hospitalization

Continuous Efflux Capacity and Cardiovascular Events

Efflux Capacity Improves Risk Prediction of ASCVD Events

Model C-statistic

Integrated Discrimination Improvement

(IDI)

Net Reclassification

Index (NRI)

Traditional Risk Factors

0.827

0.02 0.37Traditional Risk

Factors + Efflux capacity

0.841

P value/95% CI

p=0.02 p=0.0008 (0.18-0.56)

Traditional Risk Factors: age, sex, ethnicity, diabetes, hypertension, smoking, BMI, total cholesterol, log triglyceride, statin use

Summary• In contrast to HDL cholesterol, cholesterol efflux capacity (CEC)

is minimally correlated with risk factors, lipoproteins, adiposity, or insulin resistance.

• CEC is inversely associated with incident atherosclerotic CV events in a population-based study free of CVD

• This association is not attenuated by traditional risk factors, HDL cholesterol, or HDL particle concentration

• CEC, as a measure of reverse cholesterol transport, may provide the ability to interrogate key mechanisms related to cardiovascular disease in humans

Implications

Other Disease Pathways

Therapeutic Response

LDL cholesterol

Non-HDL cholesterol

HDL cholesterol

HDL Function/Flux

Risk Prediction

Coronary Disease

Free Stable ACS